Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Multicenter Study
. 2022 Sep 2;12(1):14952.
doi: 10.1038/s41598-022-18751-2.

Experimental evidence of effective human-AI collaboration in medical decision-making

Carlo Reverberi  1   2 , Tommaso Rigon  3 , Aldo Solari  4   3 , Cesare Hassan  5   6 , Paolo Cherubini  7   4   8 ; GI Genius CADx Study Group; Andrea Cherubini  9   10
Collaborators, Affiliations
Multicenter Study

Experimental evidence of effective human-AI collaboration in medical decision-making

Carlo Reverberi et al. Sci Rep. .

Abstract

Artificial Intelligence (AI) systems are precious support for decision-making, with many applications also in the medical domain. The interaction between MDs and AI enjoys a renewed interest following the increased possibilities of deep learning devices. However, we still have limited evidence-based knowledge of the context, design, and psychological mechanisms that craft an optimal human-AI collaboration. In this multicentric study, 21 endoscopists reviewed 504 videos of lesions prospectively acquired from real colonoscopies. They were asked to provide an optical diagnosis with and without the assistance of an AI support system. Endoscopists were influenced by AI ([Formula: see text]), but not erratically: they followed the AI advice more when it was correct ([Formula: see text]) than incorrect ([Formula: see text]). Endoscopists achieved this outcome through a weighted integration of their and the AI opinions, considering the case-by-case estimations of the two reliabilities. This Bayesian-like rational behavior allowed the human-AI hybrid team to outperform both agents taken alone. We discuss the features of the human-AI interaction that determined this favorable outcome.

PubMed Disclaimer

Conflict of interest statement

AC is an employee of Cosmo ai/Linkverse. CR is offering paid advice to Linkverse on a different project. The remaining authors have no conflicts of interest to disclose. We dealt with potential competing interests by preregistering our main measures, analyses, and hypotheses before data collection (https://osf.io/y9at5). Data were analyzed by TR and AS (who are free of competing interests) and are made publicly available to the scientific community (https://osf.io/57smj). Results from planned analyses, exploratory analyses, and analyses advised by Reviewers have all been reported.

Figures

Figure 1
Figure 1
Left panel: The stimuli used in the experiment were prospectively collected in a real-world clinical setting using an ai medical device supporting mds for lesion detection (cade) and categorization (cadx) as adenomatous or non-adenomatous. Right panel: An international group of endoscopists were asked to optically diagnose the same set of lesions, presented as short video clips, in two experimental sessions. In the first session (top-right panel) the ai only highlights the target lesion, while in the second session (bottom-right panel) ai also dynamically offers an optical diagnosis. For more details on the ai device see Appendix A.1.2.
Figure 2
Figure 2
md-ai team. An endoscopist subject to under-reliance discounts the added information given by the ai (a). An endoscopist subject to over-reliance supinely accepts the ai suggestion (b). The optimal use of ai should rest on an in-between, well-calibrated approach where the endoscopist uses the ai opinion for coherently revising their confidence in their initial evaluation. In this way, the medical decision-making process would benefit from a collaboration between the two intelligences (c).
Figure 3
Figure 3
Influence of the ai: alluvial diagrams representing changes in endoscopist’s opinion between the two sessions as a function of perceived ai response.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Topol EJ. High-performance medicine: The convergence of human and artificial intelligence. Nat. Med. 2019;25:44–56. doi: 10.1038/s41591-018-0300-7. - DOI - PubMed
    1. Dellermann, D. et al. The future of human–AI collaboration: A taxonomy of design knowledge for hybrid intelligence systems. arXiv:2105.03354 (2021).
    1. Akata Z, et al. A research agenda for hybrid intelligence: Augmenting human intellect with collaborative, adaptive, responsible, and explainable artificial intelligence. Computer. 2020;53:18–28. doi: 10.1109/MC.2020.2996587. - DOI
    1. Zhang, Y., Liao, Q. V. & Bellamy, R. K. Effect of confidence and explanation on accuracy and trust calibration in AI-assisted decision making. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, 295–305 (2020).
    1. Wickens CD, Clegg BA, Vieane AZ, Sebok AL. Complacency and automation bias in the use of imperfect automation. Hum. Factors. 2015;57:728–739. doi: 10.1177/0018720815581940. - DOI - PubMed

Publication types